Agricultural pollution of the environment is jointly determined by economic decisions driving land use, production practices, and stochastic biophysical processes associated with agricultural production, land and climate characteristics. It follows that environmental and economic statistics, traditionally collected independently of each other, offer little insight into non-point pollutant loadings. We argue that effective policy development would be facilitated by integrating environmental and economic data gathering, combined with simulation modelling linking economic and biophysical components. Integrated data collection links economics, land use, production methods and environmental loadings. An integrated economic/biophysical modelling framework facilitates policy analysis because monetary incentives to reduce pollution can be evaluated in the context of market costs and returns that influence land use and production activity. This allows prediction of environmental and economic outcomes from alternative policies to solve environmental problems. We highlight steps taken to merge economic and biophysical modelling for policy analysis within the Economic Research Service of the United States Department of Agriculture. An example analysis of a policy to reduce agricultural nitrogen pollution is presented, with the economic and environmental results illustrating the value of linked economic and biophysical analysis.